Vanilla+CNN+on+MNIST+dataset.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Simple CNN for MNIST \n",
    "\n",
    "Using the MNIST dataset (70 000 pictures of hand-written digits) we will train a simple CNN, which is able to predict a digit given a picture of a hand-written digit with 99% accuracy. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "np.random.seed(1337)  # for reproducibility\n",
    "from keras.datasets import mnist\n",
    "from keras.models import Sequential\n",
    "from keras.layers import Dense, Dropout, Activation, Flatten\n",
    "from keras.layers import Convolution2D, MaxPooling2D\n",
    "from keras.utils import np_utils\n",
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Network parameters:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "batch_size = 128\n",
    "nb_classes = 10\n",
    "nb_epoch = 12\n",
    "\n",
    "# input image dimensions\n",
    "img_rows, img_cols = 28, 28\n",
    "# number of convolutional filters to use\n",
    "nb_filters = 32\n",
    "# size of pooling area for max pooling\n",
    "pool_size = (2, 2)\n",
    "# convolution kernel size\n",
    "kernel_size = (3, 3)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Prepare data into training and test set."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# the data, shuffled and split between train and test sets\n",
    "(X_train, y_train), (X_test, y_test) = mnist.load_data()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "X_train = X_train.reshape(X_train.shape[0], 1, img_rows, img_cols)\n",
    "X_test = X_test.reshape(X_test.shape[0], 1, img_rows, img_cols)\n",
    "input_shape = (1, img_rows, img_cols)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "X_train shape: (60000, 1, 28, 28)\n",
      "60000 train samples\n",
      "10000 test samples\n"
     ]
    }
   ],
   "source": [
    "X_train = X_train.astype('float32')\n",
    "X_test = X_test.astype('float32')\n",
    "X_train /= 255\n",
    "X_test /= 255\n",
    "print('X_train shape:', X_train.shape)\n",
    "print(X_train.shape[0], 'train samples')\n",
    "print(X_test.shape[0], 'test samples')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# convert class vectors to binary class matrices\n",
    "Y_train = np_utils.to_categorical(y_train, nb_classes)\n",
    "Y_test = np_utils.to_categorical(y_test, nb_classes)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Build the CNN. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "model = Sequential()\n",
    "\n",
    "model.add(Convolution2D(nb_filters, kernel_size[0], kernel_size[1],\n",
    "                        border_mode='valid',\n",
    "                        input_shape=input_shape))\n",
    "model.add(Activation('relu'))\n",
    "model.add(Convolution2D(nb_filters, kernel_size[0], kernel_size[1]))\n",
    "model.add(Activation('relu'))\n",
    "model.add(MaxPooling2D(pool_size=pool_size))\n",
    "model.add(Dropout(0.25))\n",
    "\n",
    "model.add(Flatten())\n",
    "model.add(Dense(128))\n",
    "model.add(Activation('relu'))\n",
    "model.add(Dropout(0.5))\n",
    "model.add(Dense(nb_classes))\n",
    "model.add(Activation('softmax'))\n",
    "\n",
    "model.compile(loss='categorical_crossentropy',\n",
    "              optimizer='adadelta',\n",
    "              metrics=['accuracy'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Show a summary of the model parameters. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "____________________________________________________________________________________________________\n",
      "Layer (type)                     Output Shape          Param #     Connected to                     \n",
      "====================================================================================================\n",
      "convolution2d_1 (Convolution2D)  (None, 32, 26, 26)    320         convolution2d_input_1[0][0]      \n",
      "____________________________________________________________________________________________________\n",
      "activation_1 (Activation)        (None, 32, 26, 26)    0           convolution2d_1[0][0]            \n",
      "____________________________________________________________________________________________________\n",
      "convolution2d_2 (Convolution2D)  (None, 32, 24, 24)    9248        activation_1[0][0]               \n",
      "____________________________________________________________________________________________________\n",
      "activation_2 (Activation)        (None, 32, 24, 24)    0           convolution2d_2[0][0]            \n",
      "____________________________________________________________________________________________________\n",
      "maxpooling2d_1 (MaxPooling2D)    (None, 32, 12, 12)    0           activation_2[0][0]               \n",
      "____________________________________________________________________________________________________\n",
      "dropout_1 (Dropout)              (None, 32, 12, 12)    0           maxpooling2d_1[0][0]             \n",
      "____________________________________________________________________________________________________\n",
      "flatten_1 (Flatten)              (None, 4608)          0           dropout_1[0][0]                  \n",
      "____________________________________________________________________________________________________\n",
      "dense_1 (Dense)                  (None, 128)           589952      flatten_1[0][0]                  \n",
      "____________________________________________________________________________________________________\n",
      "activation_3 (Activation)        (None, 128)           0           dense_1[0][0]                    \n",
      "____________________________________________________________________________________________________\n",
      "dropout_2 (Dropout)              (None, 128)           0           activation_3[0][0]               \n",
      "____________________________________________________________________________________________________\n",
      "dense_2 (Dense)                  (None, 10)            1290        dropout_2[0][0]                  \n",
      "____________________________________________________________________________________________________\n",
      "activation_4 (Activation)        (None, 10)            0           dense_2[0][0]                    \n",
      "====================================================================================================\n",
      "Total params: 600,810\n",
      "Trainable params: 600,810\n",
      "Non-trainable params: 0\n",
      "____________________________________________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "model.summary()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "And now train the model and evaluate on the test set. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "collapsed": false,
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 60000 samples, validate on 10000 samples\n",
      "Epoch 1/12\n",
      "60000/60000 [==============================] - 115s - loss: 0.3605 - acc: 0.8891 - val_loss: 0.0885 - val_acc: 0.9721\n",
      "Epoch 2/12\n",
      "60000/60000 [==============================] - 93s - loss: 0.1356 - acc: 0.9598 - val_loss: 0.0622 - val_acc: 0.9791\n",
      "Epoch 3/12\n",
      "60000/60000 [==============================] - 94s - loss: 0.1051 - acc: 0.9685 - val_loss: 0.0534 - val_acc: 0.9825\n",
      "Epoch 4/12\n",
      "60000/60000 [==============================] - 94s - loss: 0.0880 - acc: 0.9740 - val_loss: 0.0446 - val_acc: 0.9844\n",
      "Epoch 5/12\n",
      "60000/60000 [==============================] - 93s - loss: 0.0803 - acc: 0.9765 - val_loss: 0.0399 - val_acc: 0.9864\n",
      "Epoch 6/12\n",
      "60000/60000 [==============================] - 93s - loss: 0.0720 - acc: 0.9785 - val_loss: 0.0380 - val_acc: 0.9876\n",
      "Epoch 7/12\n",
      "60000/60000 [==============================] - 93s - loss: 0.0641 - acc: 0.9805 - val_loss: 0.0366 - val_acc: 0.9879\n",
      "Epoch 8/12\n",
      "60000/60000 [==============================] - 94s - loss: 0.0591 - acc: 0.9824 - val_loss: 0.0343 - val_acc: 0.9889\n",
      "Epoch 9/12\n",
      "60000/60000 [==============================] - 94s - loss: 0.0552 - acc: 0.9834 - val_loss: 0.0330 - val_acc: 0.9879\n",
      "Epoch 10/12\n",
      "60000/60000 [==============================] - 94s - loss: 0.0520 - acc: 0.9846 - val_loss: 0.0317 - val_acc: 0.9894\n",
      "Epoch 11/12\n",
      "60000/60000 [==============================] - 94s - loss: 0.0505 - acc: 0.9847 - val_loss: 0.0320 - val_acc: 0.9898\n",
      "Epoch 12/12\n",
      "60000/60000 [==============================] - 94s - loss: 0.0471 - acc: 0.9863 - val_loss: 0.0310 - val_acc: 0.9897\n"
     ]
    }
   ],
   "source": [
    "history = model.fit(X_train, Y_train, batch_size=batch_size, nb_epoch=nb_epoch,\n",
    "          verbose=1, validation_data=(X_test, Y_test))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Test score: 0.0309755772928\n",
      "Test accuracy: 0.9897\n"
     ]
    }
   ],
   "source": [
    "score = model.evaluate(X_test, Y_test, verbose=0)\n",
    "print('Test score:', score[0])\n",
    "print('Test accuracy:', score[1])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAicAAAGHCAYAAABrpPKuAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nSeqV3wBfoC1VdAR+PwAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f60bc074ba8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(history.history['acc'])\n",
    "plt.plot(history.history['val_acc'])\n",
    "plt.title('model accuracy')\n",
    "plt.ylabel('accuracy')\n",
    "plt.xlabel('epoch')\n",
    "plt.legend(['train', 'test'], loc='upper left')\n",
    "plt.show()"
   ]
  }
 ],
 "metadata": {
  "anaconda-cloud": {},
  "kernelspec": {
   "display_name": "Python [conda root]",
   "language": "python",
   "name": "conda-root-py"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.5.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}